ISSN 0972-978X 

  About COAA








Failed Internal Fixation of Hip Fractures treated with Salvage Total Hip Arthroplasty

M Leonard, U Alao,  A Glynn, M Dolan

Department of Trauma and Orthopaedic Surgery, Cork University Hospital, Cork

Address for Correspondence:

Michael Leonard
36 An Drisin, Ballymoneen Road, Galway City, Ireland

Phone: (00353)879163747
Fax    : (00353)564675


Failed fixation of hip fractures typically leads to  functional disability and pain for the individual, technical challenges for the surgical team, and an increase in the financial burden on society. Between 1999 and 2005 41 patients (30 women and 11 men) with a mean age of 70 were treated at our institution with a total hip arthroplasty for failed dynamic hip screw fixation of a hip fracture. 
This study had three purposes: (1) to determine the reason/s for failure of internal fixation (2) to record difficulties / complications encountered in converting to a salvage arthroplasty and  (3) to compare the outcome of patients who underwent salvage arthroplasty (Group 1)  with a matched group of patients who had a primary hip arthroplasty for degenerative disease (Group 2).
Failure to achieve a good reduction and optimal screw placement was evident in 80% of cases of failed fixation. A high incidence of complications was recorded in the perioperative period during conversion to a salvage arthroplasty. Functional outcome was statistically inferior in Group1, this group also had a much higher incidence of complications. Radiographs at 2 years post operatively showed evidence of femoral stem loosening in 16% of the salvage group compared with 3% in the primary hip arthroplasty group.
Salvage hip arthroplasty is associated with higher complication rate and poorer outcome than primary hip arthroplasty. We recorded a high incidence of femoral stem loosening in  patients who had salvage hip arthroplasty, we recommend more frequent clinical and radiographic follow up of these patients.

J.Orthopaedics 2009;6(1)e4


fracture calcaneum; intraarticular; extraarticular; surgery; outcome

The incidence of hip fractures worldwide was 1.26 million in 1990 and is estimated to increase to 2.6 million in 2025 and 4.5 million in 20501. The current cost of treating these injuries is estimated to be 10 billion dollars per year in the United States alone2. Failed surgical treatment of hip fractures typically leads to profound functional disability and pain for the individual, technical challenges for the surgical team, and an increase in the financial burden on society. 
Epidemiological studies have demonstrated that the anatomical location of the ‘hip fracture’ is split evenly between the  neck of femur and the intertrochanteric region3. In both types the fracture pattern, bone quality, accuracy  of reduction and adequacy of fixation are key factors determining outcome4. There are important differences however between these two sites ; a fracture of the femoral neck may cause irreparable damage to the blood supply of the femoral head  leading to avascular necrosis and femoral head collapse whereas an intertrochanteric fracture is subject to significant biomechanical loads which can lead to loss of reduction and/or fixation failure. 
Treatment options for failed internal fixation consist of non-operative care for the very elderly or medically unfit patient, revision internal fixation, or hip arthroplasty. Revision internal fixation has the benefit of retaining the native joint which may be beneficial in the physiologically young patient. Frequently, however the patient has poor bone quality, an unfavourable fracture pattern, damaged femoral head, damaged articular cartilage and limb shortening. Due to these issues arthroplasty, with excision of head-neck fragment is an accepted salvage technique. The technical challenges of performing a salvage arthroplasty in this situation include the presence of failed internal fixation devices, difficulties with dislocation,intra-operative fracture, bone deformity, cement extravasation, leg length discrepancy, stress risers and poor bone quality. 
This study had three purposes: (1) to determine the reason/s for failure of internal fixation  in a group of patients treated in our institute for failed internal fixation of a proximal femoral fracture(2) to review the conversion in these patients of failed fixation to total hip arthroplasty and record any difficulties / complications encountered, and (3) to compare the outcome of patients who underwent primary hip arthroplasty for degenerative arthritis with those patients in this study who underwent conversion of failed fixation to arthroplasty

Patients and Methods:

Between 1999 and 2005 41 patients (30 women and 11 men) with a mean age of 70 were treated at our institution with a total hip arthroplasty after failed internal fixation of a hip fracture; 24 neck of femur, 17 intertrochanteric fractures. All were initially  fixed with a dynamic hip screw (DHS) (AO Synthes, Switzerland) construct and had undergone revision to arthroplasty for a combination of reasons including pain, stiffness, and decreased mobility coupled with radiographic evidence of a complication of the initial fracture treatment.

Patients were identified from our institutions computerised database.

The radiographs and medical charts of all patients were obtained following institutional approval.

Intracapsular femoral neck fractures were classified according to the Garden system5 and intertrochanteric fractures were classified as stable or unstable as per Kyle et al6.

The time lag from fracture to definitive reduction and fixation for intracapsular fractures was recorded from the medical notes.

The quality of the reduction of the fracture achieved was assessed on the basis of displacement and alignment of the fracture. The reduction was categorized as good, acceptable, or poor. For a reduction to be considered good, there had to be normal or slight valgus alignment on the anteroposterior (AP) radiograph, less than 20 degrees’ angulation on the lateral radiograph, and no more than 4mm of displacement of any fragment. To be considered acceptable, a reduction had to meet the criterion of a good reduction with respect to either alignment or displacement or both. A poor reduction met neither criterion7.

The technical quality of the radiographs available and the patient positioning were too inconsistent  to allow for quantification of osteoporosis with the use of the method of Singh et al.

The tip-apex distance (TAD), the sum of the distance from the tip of the lag screw to the apex of the femoral head as described by Baumgaertner et al.was used to describe the position of the screw in the femoral head. They demonstrated that a TAD of 25 millimeters (mm) or less resulted in no cut-outs, a TAD of 35 – 45mm resulted in a cut-out rate of 36% and a TAD of over 45mm a cut-out rate of 60%7.

The time to revision was recorded as the length of time in months between initial fracture fixation and salvage arthroplasty.

Each patient who had undergone salvage arthroplasty (Group 1) was matched with a patient who had undergone total hip arthroplasty for degenerative disease in our unit (Group 2). Patients were matched for age, sex, implant and time since insertion of the implant. The  vast majority  (>95%)of patients in both groups had a cemented total hip  with either an Exeter or a Charnley stem in combination with a polyethelene cemented cup. The patients in Group  2 were selected without any knowledge of their outcome to eliminate selection bias.

All surviving patients form both groups were followed up for a minimum of two years (mean 5 years). Three main outcome measures were compared between the two groups; surgical complications,  the Oxford hip score( interpreted as per Murray et al. with a continuous score ranging from 0 (most severe symptoms) to 48 (least symptoms / best outcome))8, and radiographic analysis of the femoral component for signs of loosening. Criteria of loosening were defined (Table 1), and the standard 2 year post-operative radiographs were compared with those obtained in the early (1-3days) postoperative period.

Statistical Methods

Statistical analysis were performed with SPSS 13 (SPSS, Chicago, Illinois). Between group comparisons were made using Mann-Whitney U tests. A P- value of less than 0.05 was considered to be significant. 

Progressive lucent zone at bone-cement interface >2mm

Subsidence >5mm

Varus displacement

Fracture of the cement

Appearance of a lucent zone between metal and cement/bone

Table 1. Radiographic signs of loosening of the femoral stem


Details of the patients are summarised in Table 2. Mean follow up was 5 years (range 9 – 2.5 yrs).In the 11 unstable intertrochanteric fractures, failure of fixation was due to a combination of both inadequate reduction (acceptable reduction in 6, poor reduction in 5) and screw placement (<35 in 8, <45 in 3). The combination of these factors lead to a non-union in 5, cut-out in 3 and implant breakage in 3 (Figure 1).  Of note all three cases of implant breakage were associated with a short screw; tip-apex distance (TAD) <45 mm). 

Patients             Fracture               Reduction             TAD         Cause of failure         Time to revision
No. Age. Sex

1     79    M        Intertroch. – stable      Acceptable    <35mm        Cut-out      16months

2     85    M        Intertroch – unstable    Acceptable    <35mm        Cut-out       2months

3     55    M        Intertroch – unstable     Poor              <35mm        Non-union   6 months

4     88    M        Neck of femur- G1         Good             <25mm         AVN            15months

5    76     F         Neck of femur- G2         Good             <45mm        Cut-out        14months

6    76     F         Intertroch - unstable     Acceptable    < 35mm       Non –union  14months

7    73     F         Neck of femur – G1        Good             <45mm        Cut-out       36months

8    85     F         Neck of femur – G2        Good             <25mm        AVN             36months

9    84     F         Neck of femur – G3        Good             <35mm        AVN              12months

10  65     F         Neck of femur – G2        Good             <25mm        AVN              10months

11  51     F         Neck of femur – G1        Good             < 35mm       AVN               18months

12  84     F         Intertroch – unstable    Acceptable    <45mm   Implant breakage13months

13  78     F         Intertroch – stable         Poor              < 35mm    Cut-out           29months

14  66     F         Intertroch – unstable     Acceptable    <45mm   Implant breakage37months

15  70     F         Intertroch – stable         Good              <45mm       Cut-out     10months

16  56     M       Neck of femur – G3         Acceptable     <25mm       AVN           31months

17  80     F         Intertroch – unstable     Poor               <35mm      Cut-out      4months

18  75     F         Neck of femur – G2        Good               <45mm      Cut-out     1month

19  79     F         Neck of femur – G2       Good               <35mm      Cut-out      3months

20  70     F         Neck of femur – G2       Good               <25mm      AVN           38months

21  73     F         Neck of femur – G3       Acceptable     <35mm     Non-union   37months

22  54     M       Intertroch – unstable     Acceptable      <35mm     Non-union  19months

23  62     M       Neck of femur – G4       Acceptable      <25mm     AVN             18months

24  53     F        Neck of femur – G3       Acceptable       <35mm    Non-union   12months

25  80     M       Neck of femur – G1       Good                <25mm     AVN            18months

26  58     F        Neck of femur – G3       Acceptable       < 35mm    AVN            9months

27  67     M       Neck of femur – G3      Good                < 25mm    AVN              18months

28  80     F        Neck of femur – G1       Good                <45mm     Cut-out        1month

29  80     F        Intertroch – unstable    Acceptable      <35mm     Non-union     11months

30  81     F        Intertroch – stable        Acceptable       <45mm   Cut-out         2months

31  84     F        Intertroch – unstable    Poor       <45mm    Implant breakage  21months

32  56     F        Neck of femur – G3       Acceptable        <35mm    AVN            16months

33  90     F        Neck of femur – G1      Good                 <25mm    Cut-out        15months

34  70     F        Neck of femur – G1      Good                 <35mm    AVN             12months

35  82     F        Intertroch – stable      Acceptable       <35mm    Cut-out          3months

36  89     F        Neck of femur – G1      Good                <25mm    AVN               5months

37  64    M       Neck of femur – G3        Acceptable       <35mm   Cut-out          1month

38  66    F        Intertroch – unstable     Poor                  <35mm    Non-union   12months

39  68    M       Intertroch – stable         Acceptable       <35mm    Cut-out        16months

40  66    F        Intertroch – unstable      Poor                 <35mm     Cut-out      16months

41  57    F        Neck of femur – G3      Acceptable        <25mm    AVN               24months

Intertroch = Intertrochanteric,  G1 = Garden 1(incomplete valgus impacted intracapsular fracture neck of femur), G2 = Garden 2 (complete fracture, no displacement), G3 = Garden 3 (Complete fracture, partial displacement), G4 = Garden 4 (complete displacement)  TAD = Tip-Apex Distance, AVN = Avascular Necrosis

Table 2. Summary of patients’data

Figure 1: Failed fixation of an intertrochanteric fracture with implant breakage

In the 6 stable intertrochanteric fractures , an acceptable reduction was obtained in 4, with 1 good and 1 poor reduction. However in no case was the ideal TAD achievied, 4 had a TAD of <35 and 2 a TAD of <45, all failed due to cut-out of the screw.

Of the 14 intracapsular undisplaced (Garden 1 + 2) neck of femur fractures all were classified as having a good reduction. Failure occurred due to cut-out in 6 cases (Figure 2), 5 of which had inadequate screw length ( TAD <45mm in 4 and <35mm in 1). The remainder failed due to the development of avascular necrosis (AVN).

Figure 2: Superior cut out of screw in an intracapsular fracture

In the displaced neck of femur fractures (Garden 3+ 4) a good reduction was obtained in 2 and an acceptable reduction in the remaining 8. There were no open reductions. Screw placement was also satisfactory in this group; TAD <25mm in 4 and <35mm in 6. Failure occurred due to the development of AVN in 7, non-union in 2 and cut-out in 1.

The mean time from fracture to fixation for the patients who developed AVN was 21.1 hours (hrs) compared to 16.2 hrs in the non AVN group, however this was not significant (p =0.164).

The mean time between fracture fixation and subsequent revision to arthroplasty was 16 months (range 1 – 38 months).

On review of the operative notes of the 41 cases, the mean surgical time was 2 hours and 55 minutes. Mean blood loss was 923millilitres. Difficulty with dislocation was documented on 11 occasions, difficulty with removal of metal was documented on 10 occasions. Cement extrusion through previous screw holes was noted on 11 post-operative radiographs. One patient had two episodes of hip dislocation while still an in-patient, these were treated successfully with closed reduction (Figure 3). Two patients required re-operation, one for debridement of a sterile wound haematoma, and another for rewiring of a trochanteric nonunion. On one occasion only was the surgeon forced to change the operative plan from a cemented to a uncemented long stemmed implant with cables due to both a fractured greater trochanter and difficulty with removal of the previous metal work.

Figure 3:  Dislocated salvage total hip replacement 2 days post-operativley

Before the follow-up date 10 of the 41 patients had died. All were over 80 years of age at the time of their death. Two died within 1 year of revision of their fixation. One patient died within 1 month due to a deterioration in her extensive medical co-morbidities, although radiographs were satisfactory, she had never ambulated post-operativley. Another patient died within 6 months of surgery, again from a combination of medical problems, she had been progressing on a walking frame prior to his deterioration. The remaining 8 patients died of illnesses unrelated to their surgery at a later stage.

All of the 41 patients who had undergone revision of their fixation to total hip arthroplasty (Group1) were matched with a group who had undergone primary hip arthroplasty for degenerative disease (Group2) as described in the methods section. At the follow up date 4 of the patients in Group 2 had died, none within 1 year of their surgery. No deaths in this group were attributable to their hip surgery.

Of the remaining 31 patients in Group 1 we were unable to contact 7 leaving  24 patients available for follow up in Group 1. Of the remaining 37 patients in Group 2 we were unable to contact 4 patients and 4 did not wish to participate in the study, leaving 29 patients for follow-up in Group 2.

All of the patients in Group 1 were initially asked if conversion to a total hip replacement had (a) provided pain relief and (b) improved their ability to ambulate; 23 out of the 24 patients felt that the surgery had been successfully in achieving these aims.

Details of the follow-up of patients in Group1 and 2 are summarised in Table 3. The overall incidence of complications was much higher in Group 1 (16) compared with Group 2 (5). The mean oxford hip score for Group 1 was 30 compared with 43 for Group 2. Comparative analysis showed this to be significant (p=0.03). 

Group Group2

No. of Patients                                 24                      29
Median Age                                     75                      76
M:F ratio                                       7:17                  10:19

Surgical Complications

Haematoma                                    3                         1
Superficial Infection                         6                         2
Dislocation                                      3                         0
DVT                                                2                         2
Re-operation for any                        2                         0

Mean Oxford hip score 
at mean follow up of  5 years
           30                         43

Table 3. Surgical Complications and Oxford Hip Score (OHS) in the two groups

A 2year postoperative radiograph was available for comparison with the early postoperative radiograph in 31 of the patients in Group1 and 33 of the patients in Group 2. As defined by the criteria in Table 1 there was evidence of loosening of the femoral stem in 16% (5/31) of the patients in Group 1; 3 with progressive lucent zones at the cement bone interface, and 2 with varus displacement and subsidence. Only 1 patient in Group 2 showed evidence of loosening with a progressive lucent zone at the cement bone interface.


A number of factors may impact negatively on the outcome of internal fixation of hip fractures. The two factors which are under the surgeons direct control however are; the quality of the reduction, and the accuracy of insertion of the fixation device.

Of the 17 cases of failed fixation of intertrochanteric fractures treated in this study none had the ideal combination of good reduction and optimal location of the screw in the femoral head. Union rates of up to 100% have been reported in well-reduced, intertrochanteric  fractures treated with ideal implant placement9, however failure rates of up to 56% have also been reported when this optimal situation is not achievied4. This problem has led to the design of several types of fixation devices including intramedullary devices, however none have shown a clear clinical advantage over the dynamic hip screw (DHS), which has a definite cost-benefit advantage. To date no single implant is universally accepted for the treatment of these fractures.

The optimal surgical treatment of intracapsular femoral neck fractures in adults and eldery patients remains controversial. The two options are prosthetic replacement or internal fixation. Proponents of prosthetic replacement argue that replacement of the femoral head eliminates the necessity for revision surgery due to avascular necrosis (AVN) and non-union10. Those in favour of internal fixation report decreased operative time, blood loss, and mortality rates11. Of the 24 femoral neck fractures in this study failure due to screw cut-out with a suboptimal placement occurred in 6 patients. Failure due to AVN occurred in 15 cases in spite of all having had a satisfactory reduction and fixation. The literature would suggest that the development of AVN in such cases is primarily due to the degree of initial displacement, and the length of time between fracture and fixation10, 11. 

When converting failed internal fixation of a hip to an arthroplasty, a number of technical challenges must be overcome. Difficulties with dislocation and intraoperative fracture have been reported12.The internal fixation device which has failed, often has broken screws which must be removed, also the ununited head-and neck fragment or fragments are usually in a deformed position and must be mobilised before being excised. This requires additional dissection placing nearby neurovascular structures and muscles at risk and leading to increased blood loss12.

It has been suggested that the results of salvage total hip arthroplasty following failed internal fixation  are comparable with those of primary joint replacement12,13. The patients in our study who required a salvage arthroplasty had a greater prevalence of complications and poorer functional results than did patients who underwent primary hip arthroplasty despite the use of similar techniques and implants. The reasons for these differences are probably multifactorial. Patients with failed internal fixation often have a prolonged period of hip pain and immobility leading to muscle wasting and disuse osteoporosis prior to revision surgery. The increased technical challenges described earlier, prolonged operating time, increased exposure, altered anatomy and greater blood loss all are likely to play a role in poorer outcome. However conversion to hip arthroplasty alleviated pain and improved function in the vast majority of these patients, which is the hallmark of an effective salvage procedure.

All cases in this study were performed by experienced hip surgeons using modern techniques and implants, in spite of this however we still recorded radiographic evidence of femoral stem loosing at 2 years  in 16% of the patients who had undergone conversion to arthroplasty. It has been suggested that cortical holes left by previous screws may lead to cement extravasation with suboptimal pressurisation, and poor remodelling of the cortical bone, leading to an inferior cement mantle and potential stress risers at areas of cement extrusion14. Is it intuitive that every effort should be made to obtain the best mantle possible but we would also recommend that these patients be followed up for longer periods with more frequent radiographs than the standard hip replacement patient.

 In summary when undertaking surgical stabilisation any hip fracture one should make every effort to achieve the best reduction and most accurate fixation possible, failure to achieve these goals was evident in 80% (33/41) of our cases. Factors such as osteoporosis, compliance with post-operative mobilisation and delay in fracture fixation are to some extent ‘out of the surgeons hands’. Conversion to arthroplasty is technically challenging, and is associated with higher complication rate and poorer outcome than primary hip arthroplasty. We recorded a high incidence of femoral stem loosening in patients who had undergone conversion to hip arthroplasty for failed fixation, as a result we recommend more frequent clinical and radiographic follow up of these patients.

Reference :

  1. Lorich DG, Geller DS, Nielson JH Osteoporotic Pertrochanteric Hip Fractures: Management and Current Controversies. J Bone Joint Surg Am 2004; 86 2:398-410

  2. Sattin RW Falls among older persons: a public health perspective. Annu Rev Public Health 1992;13:489-508

  3. Apple DF, Hayes WC. Prevention of falls and hip fractures in the elderly. Rosemont, IL: American Academy of Orthopaedic Surgeons 1993

  4. Haidukewych GJ, Israel TA, Berry DJ  Reverse obliquity fractures of the intertrochanteric region of the femur. J Bone Joint Surg Am 2001; 83:643-50

  5. Garden RS  Low-angle fixation in fractures of the femoral neck.J Bone Joint Surg Br  1961;43-B:647-63

  6. Kyle RF, Gustilo RB, Premer RF Analysis of six hundred and twenty-two intertrochanteric hip fractures. A retrospective and prospective study. J Bone Joint Surg Am 1979;61:216-221

  7. Baumgaertner MR, Curtin SL, Lindskog DM  The value of the tip-apex distance in predicting failure of fixation of peritrochanteric fractures of the hip. J Bone Joint Surg Am 1995; 77:1058-1064

  8. Murray DW, Fitzpatrick R, Rogers K.  The use of the Oxford hip and knee scores. J Bone Joint Surg Br 2007;8:1010-14

  9. Baumgaertner MR, Solberg BD  Awareness of the tip-apex distance reduces failure of fixation of trochanteric fractures of the hip. J Bone Joint Surg Br  1997;79:969-71

  10. Chua D, Jagial SB, Schatzker J.  An orthopaedic surgeon survey on the treatment of displaced femoral neck fracture: opposing views. Can J Surg 1997;40:271-7

  11. Parker MJ, Pryor GA. Internal fixation or arthroplasty for displaced cervical hip fractures in the elderly: a randomised controlled trial of 208 patients. Acta Orthop Scand 2000;71:440-6

  12. Tabsh I, Waddell JP, Morton J. Total hip arthroplasty following failed internal fixation of hip fractures. Clin Orthop 1997;11:166

  13. Mehlhoff T, Landon GC, Tullos HS. Total hip arthroplasty following failed internal fixation of hip fractures. Clin Orthop 1991;269:32-7

  14. Patterson BM, Salvati E, Huo MH.  Total hip arthroplasty for complications of intertrochanteric fracture. J Bone Joint Surg Am 1990;72: 776-77


This is a peer reviewed paper 

Please cite as: M Leonard: Failed Internal Fixation of Hip Fractures treated with Salvage Total Hip Arthroplasty

J.Orthopaedics 2009;6(1)e4





Arthrocon 2011

Refresher Course in Hip Arthroplasty

13th March,  2011

At Malabar Palace,
Calicut, Kerala, India

Download Registration Form

For Details
Dr Anwar Marthya,
Ph:+91 9961303044



Powered by



Copyright of articles belongs to the respective authors unless otherwise specified.Verbatim copying, redistribution and storage of this article permitted provided no restrictions are imposed on the access and a hyperlink to the original article in Journal of Orthopaedics maintained. All opinion stated are exclusively that of the author(s).
Journal of Orthopaedics upholds the policy of Open Access to Scientific literature.